PROJECT SUMMARY/ABSTRACT Suicide is a major public health concern in adolescents and the second leading cause of death in this age group, highlighting the need to investigate suicide in adolescents. The two most prominent risk factors for completed suicide in youth are a past suicide attempt and a diagnosis of a depressive episode, each independently representing a 10- to 30-fold increased risk for completed suicide. However, while over 80% of suicidal youth present with depressive symptoms at the time of the attempt, only 30% of depressed youth will ever attempt suicide. These data suggest that depressed adolescents with prior suicide attempts represent a distinct neurobiological subgroup. However, at present, we have no biomarkers to identify those depressed youth who will engage in suicidal behavior. This proposal addresses this concern. Our proposed model is: (1) anhedonia is associated with heightened suicidal risk in depressed youth; (2) both positive valence (PVS) deficits (reward expectancy, attainment, positive prediction error) and negative valence (NVS) hyperactivity (negative prediction error, pain avoidance) contribute to anhedonia and underlie suicidal behavior in depressed youth; (3) the habenula (Hb), a small limbic hub, plays a key role in PVS/NVS processes by inhibiting reward signaling in response to pain and loss; and (4) recent advances in high-resolution MRI address prior technical constraints in studying small structures such as the Hb. These advances now allow us to study the neuronal circuitry underlying PVS and NVS in a neuroanatomically rigorous framework. In support, we documented that of adolescent depression?s core symptoms, only anhedonia, not irritability, was associated with suicidality among depressed youth. Similarly, we found that both anhedonia and entrapment were independently associated with suicidality in acutely suicidal hospitalized individuals. We also developed the reward flanker (RFT) and reward prediction error (RPET) fMRI tasks and identified brain function during reward anticipation, attainment, and positive and negative prediction errors. Notably, both RPET and a thermal pain task elicited measurable Hb activation. Additionally, utilizing an optimized segmentation and seeding approach, we mapped Hb intrinsic functional connectivity in adults and in adolescents with psychiatric symptoms. Building upon our work, we propose a tightly integrative study to test the overall hypothesis that both PVS deficiency and NVS hyperactivity underlie suicidal behavior in depressed youth. We will study 30 suicidal and 30 non-suicidal depressed adolescents, ages 12-17, group matched for age, sex, and handedness. All will have comprehensive clinical evaluations (diagnoses, behavioral measures, computerized reward task). Neuroimaging studies will include structural MRI, resting-state and task fMRI (RFT, RPET, thermal pain). Machine learning algorithms will be used to examine which PVS/NVS processes predict suicidal behavior in depressed youth.